Computer Vision - COVID19 Image Classification. - By David Salako.


Background, Goal, and Context.

Provided are a training set and a test set of images of 3 classes. Each image has a filename that is its unique id. The dataset comprises 3 classes:

The goal of the project is to create a classifier capable of determining the class of the X-ray image.

The context is:


Dataset:

The project is inspired from a dataset from Kaggle. Note: For project purposes, the dataset is modified according to the learning requirements. You are provided with the following files:

The dataset (above 4 files) was downloaded from the Olympus platform of Great Learning.


Objective.

To implement the techniques learned as a part of the course with the following learning outcomes:


Solution:

Importing the libraries.

Load the Dataset.

Shape of the Data.

Observations:

Observations:

Observations:

Label Data Exploration.

Observations:

Observations:

Observation:

Image Data Exploration.

Image Dimensions

Are all images the same size?

Visualize Images.

Observation:

Visualize the Gray Scale X-ray Images.

Observations:

Visualize the Black and White RGB X-ray Images.

Mean per each Diagnosis.

Observations:

Data Pre-Processing.

No Gaussian Blurring et al.

Observation:

Gaussian Blurring.

Observation:

Normalization.

Observations:

Data Preparation.

Labels to one-hot vectors.

Split Original Unblurred Data (Train, Test, and Validation).

Split Blurred Data (Train, Test, and Validation).

Check Data Compatibility with Keras.

Model Building.

Helper Functions

Model1 - Basic Model.

Build and Train Model.

Plot Train/Validation Metrics.

Plot Confusion Matrix.

Model1 Observations:

Model2 - Model with more layers and Dropout.

Build and Train Model.

Plot Train/Validation Metrics

Plot Confusion Matrix

Model2 Observations:

Model3 - Model with Dropout,DataAugumentaion, LRReduction Callbacks to address the overfitting.

Build and Train Model.

Plot Train/Validation Metrics.

Plot Confusion Matrix.

Model3 Observations:

Model4 - Basic Model (Original Dataset - No Gaussian Blurring et al.).

Build and Train the Model.

Plot Train/Validation Metrics.

Plot Confusion Matrix.

Model4 Observations.

Visualize Predictions.

Visualize predictions for x_test[2], x_test[3], x_test[8], x_test[13], x_test[25].

Model1 - Basic Model (With Gaussian Blur).

Model2 - Model with more layers and Dropout.

Model3 - Model with Dropout, DataAugumentation, LRReduction Callbacks to address the overfitting.

Model4 - Model with Original Dataset and No Gaussian Blurring et al..

Observations:

Model Performance Evaluation.

Model Performance Comparision.

Conclusion and Key Takeaways.